Column Type Annotation

Column Type Annotation (CTA) is a sub-task of table annotation that aims to predict the semantic type of a column in a table. This task involves selecting appropriate semantic labels for table columns from vocabularies such as DBpedia, Schema.org, or WikiData. CTA can be treated as a multi-class classification problem, where only one semantic type is annotated for each column, or as a multi-label classification problem, allowing a column to have multiple semantic types. This task is of significant application value for enhancing data understanding, promoting information retrieval, and strengthening knowledge graph construction.